Efficient KNN search through high-dimensional indexing with iDistance

by Mike Schuh for PostgreSQL Project

I propose to implement the state-of-the-art multi-dimensional indexing algorithm iDistance directly in PostgreSQL to better facilitate fast and efficient exact KNN search in high-dimensional data. This is an ideal follow up to a project last year (by another student) on efficient range queries, and showcases the community-oriented response of PostgreSQL towards modern trends in research and application.